Separation Mastering?

[Mattierocks]

I was reading some online articles about mastering and came across a mastering house that talks about "separation mastering." Anyone have any experience with this? Thoughts on this particular technique? Thanks!

[kylen]

Sounds interesting, What's the link?

[chessrock]

The more common terminology for this is called "stems."

And it's somewhat common, most often sending the ME separate vocal "stems" and music "stems."

Something like this would just basically allow the ME a little more control over the final outcome (i.e. relative volume of the vocal track, etc.)

.

[nevome]

I'm pretty sure this is the process in which you do not hand a single stereo track to the mastering engineer but rather hand them several stereo tracks taken from your mix by muting and soloing various tracks and either bouncing them to disk or or recording new stereo tracks from your DAW in real time by routing the output of your converter back into a pair of its inputs.

The 'separations' or groups [I find my self mixing a very similar method in Cubase with group channels] are clusters of like instruments ie you would send in maybee 5 for a rock mix, *drums*electric guitars*bass*backup vocals*lead vocals.

The process allows the mastering engineer to apply eq and compression [possibly even multiband or M/S processing] to each set of instruments without mucking with the others and I have read that it allows for a clearer and 'deeper' audio image.

I plan on trying it with a band I'll be recording for here in the next few weeks and I plan on A/B comparing it to my old habbit of simply erasing below 35Hz, boost highs if it lacks and give a few dB of limiting compression.

While there is no replacement for a good mix, I hope to see that even novices can put this method to use with good results.

EDIT : or since its reported as a waste, maybee I wont bother

[Mattierocks]

That's correct. It looks like you give them the whole stereo mix, then mixes for each instrument/vocal/etc. Apparently this gives the mastering engineer the flexibility to adjust each component individually. Here's the link:

Separation Mastering

[Mattierocks]

And... I just noticed this article which I overlooked before...

EQ Article on separations

[bblackwood]

It's vestman's attempt to sound like he's done something original. Stems have been around since the 60's.

I consider it mixing in the mastering room and think it's a bad idea, personally...

[Mattierocks]

Brad, I checked out your website... What did you master for Sister Hazel?

[chessrock]

Originally Posted by bblackwood

I consider it mixing in the mastering room and think it's a bad idea, personally...

Yup. With the exception of a few rare instances, but generally speaking I agree.

.

[bblackwood]

Originally Posted by Mattierocks

Brad, I checked out your website... What did you master for Sister Hazel?

The studio album 'Lift', the live album 'Live Live', and both the surround and stereo portions of the 'A Life in the Day' DVD.

[Mattierocks]

Originally Posted by bblackwood

The studio album 'Lift', the live album 'Live Live', and both the surround and stereo portions of the 'A Life in the Day' DVD.

Awesome! Of course, I have everyone of those and love them... I actually was listening to Lift in my truck on the way to work this morning!

[masteringhouse]

Originally Posted by bblackwood

It's vestman's attempt to sound like he's done something original. Stems have been around since the 60's.

I consider it mixing in the mastering room and think it's a bad idea, personally...

Brad is correct about the term "separation mastering". It has seems to have even found it's way into Wikipedia:

http://en.wikipedia.org/wiki/Mastering

Personally I find that there can be some situations where there are advantages in using stems. Brad and I have debated this elsewhere here, so I won't repeat all of that. In general I find that stems can be useful when the mix isn't quite up to par, there are problems in the acoustical environment where the mix is being done, when people are not quite sure what the mix will sound like after mastering and want to prevent having to create multiple vocal/bass/drum up/down mixes, and a few other situations such as an alternative to having to use M/S processing (in some situations) and when radio edits need to be done.

If a mix is good to great however, and no additional edits to the individual tracks need to be done after mixing, stems are superfluous.

[Massive Master]

It's certainly a handy thing for those reasons listed and others - Instrumental versions for live or soundtrack use, and the like -

I had a project a couple weeks ago that needed an "all normal," a no vocal, a no rhythm section and a... Something else... I remember there were four...

Anyway - Obviously, it's about the only way to efficiently handle something like that. But 90% of the time, if it's a "standard" job, I'd much rather work from "standard" mixes.

[bryank]

i dont think i would benifit from the sub-group stem mixing/mastering. i can under stand if ther are ALOT of instuments involved.....maybe over 10-12 tracks of stuff or more......maybe then it would be a good idea to try this type of mix/mastering. the more tracks you have, for me the harder it seems to mix. Sub-groups then would sound like a good idea. the less tracks you have.....mixing is a piece of cake.

but for the most part, i keep my tracks to a minimum.

[SouthSIDE Glen]

Originally Posted by bryank

i can under stand if ther are ALOT of instuments involved.....maybe over 10-12 tracks of stuff or more......maybe then it would be a good idea to try this type of mix/mastering. the more tracks you have, for me the harder it seems to mix.

Agreed.

The question is, however, where the mixing leaves off and the mastering begins. If one is doing everything himself, then submixing before doing the final 2-mix is often a good idea. In that case, though, the submixes are not really "stems", they are just submixes as part of the overall mixing process; e.g. sub mix the rhythm section first, then submix the gits and keyboards, then submix all the vocals. Then you have just three submix groups to mix together into the stereo mixdown. But then you stll have the stereo mixdown to performing the mastering process to as always.

With stems - or seperation mastering (BTW, does that mean that Brad has "seperation anxiety"? ) - one is not really making a 2-mix or stereo mixdown (except maybe as an example mix for the ME to use as a guide), but rather the mixing engineer is creating a series of submixes and leaving the final mixing for the ME to do.

I agree with John that if there's a need to create different final mixes for multiple release from the studio, that stems are a good idea (even required in some cases). But if one is just creating a single stereo product, then the mix engineer should, IMHO, create the mix himself and leave the ME just to polish the mix and perform all the other standard mastering duties. For me, asking the ME to do the mixing is asking the ME to do my job for me, which means that I am either just phoning in my work and should be fired for it, or I just don't have the chops to be a mixing engineer and should be fired for it.

If one is doing both the mixing and mastering themselves, and they submix first, and then turn to seperation mastering by mastering the submixes first before doing the final mixdown, that just seems like putting the cart before the horse to me. The whole idea of mastering processing is to polish the mix; how can one polish the mix properly before it's even assembled (unless one is really, really, really good )?

It's like assembling the fenders, doors and quarterpanels of a car; one does not polish them individually and then put them together uness they want that polish job to get all smeared and fingereprinted and generally messed up. They assemble the pieces first and then apply the polish as the final step. It seems to me that - with the special purpose exceptions made by John and Tom - mastering should be treated the same way; it is the polish that needs to be applied after the mix is completed, not while the mix is still under construction.

IMHO YMMV TGIF QWERTY

G.

[noisewreck]

To me, this whole thing sounds way too much like "separation anxiety".

[bryank]

exactly.........what he just said!

for my recordings, ill mix and master it myself. BUT i dont think i would want stems. i feel as if i have more control over ach individual track rather than working with a group of tracks to mix.

plus like what Glen just said is very true.......sending stems to a ME is like asking him to mix it too.......and master it! ME is supposed to give it the final touch and polish the over all stereo track. Sending him/her stems.......the ME is most likely gonna go ..." great......more work"

[Chip Evans]

In my wacky world of caveman engineering a "stem" would never be in the hands of mastering engineer. Unless, the client needed help and was asking for mixing assistance. An engineer, if he was doing the final master, would have very specific instructions from the producer or chief engineer. They are not subject to discussion. As I am the mastering engineer, there is usually no problem.

I guess the next thing will be a MME, Master Mix Engineer. Indeed, if
a "track" can be a "stem" then any thing can happen in pixie dust land. I do see this trend of lines blurring between specialties as a wonderful "value added" product to any professional audio engineer. Beware though, wait till a desk jockey gets a hold of the ME's final, perfect master to mix and gets paid for it.

Good thread all.

Chip Evans

[bblackwood]

I think it worth noting that not one of my major label clients (the ones typically printing various mixes for release in different formats or edits for various performances/radio) send stems - they simply print the versions needed during mix, occasionally printing edit pieces if a particular edit will be tough to achieve without it.

Why is that?

Easy - because a mix is a performance and often that performance contains buss processing. In fact, I'd wager better than 90% of the mixes that come in have buss compression. Buss compression = no stems.

I just think it's wise for the mixer to really learn their craft, not relying on the mastering engineer to finish their job.

And think about it - stems mean more money in my pocket as they invariably take longer - my motivation for this stance is pure. I've seen too many engineers become reliant and stagnate instead of really blossoming because of stuff like this.

Another one to consider - I'll bet dollars to doughnuts none of the great records you love were done with stems (and let's not forget the technology has been there for well over 40 years).

Stems are not new and not good, imo.

[masteringhouse]

Originally Posted by bblackwood

Easy - because a mix is a performance and often that performance contains buss processing. In fact, I'd wager better than 90% of the mixes that come in have buss compression. Buss compression = no stems.

Brad,

Could you explain this again, I had forgot your reasoning in regards to buss processing.

If one has a Pro Tools session with stems, and inserts an SSL (or whatever) compressor on the master bus during mastering, how is this different than doing it while mixing?

I think that I would argue that the summing on an SSL or other large format console is going to beat doing this in the box over the compression issue. This would be another reason to not use stems. Of course if you're using a cheap mixer, or mixing in the box it's irrelevant again.

Personally I don't care (nor do I charge more for stems), whatever is best for the product. I've had some mixes that were dramatically improved by using stems, but I agree that most of my higher end clients don't require this. Recommending separation mastering before hearing the stereo version (or knowing the engineer) to determine if it required is a suspicious practice.

[bblackwood]

Originally Posted by masteringhouse

If one has a Pro Tools session with stems, and inserts an SSL (or whatever) compressor on the master bus during mastering, how is this different than doing it while mixing?

Few mixers (that I'm aware of) strap a compressor across the buss at the end of the mix - they do so early on and mix to it. So unless you're talking about plug-in compressors (which few mastering houses use), the mixer would have to send his hardware unit along with notes to replicate the buss compression...

[masteringhouse]

Originally Posted by bblackwood

Few mixers (that I'm aware of) strap a compressor across the buss at the end of the mix - they do so early on and mix to it. So unless you're talking about plug-in compressors (which few mastering houses use), the mixer would have to send his hardware unit along with notes to replicate the buss compression...

Not talking about plug-in compressors, one can insert a hardware comp just as easily as a plug-in in PT. Yes, it's going through another D/A conversion, but very likely it's going to anyway.

There's also no reason why the engineer can't mix this way and send the group outputs to create the stems sans compression along with the stereo mix for comparison. If the ME has an SSL comp. (or most likely one that's even better), the ratios/attack/etc. should be able to be reproduced without a major problem.

[bblackwood]

Originally Posted by masteringhouse

There's also no reason why the engineer can't mix this way and send the group outputs to create the stems sans compression along with the stereo mix for comparison. If the ME has an SSL comp. (or most likely one that's even better), the ratios/attack/etc. should be able to be reproduced without a major problem.

Come on, Tom - you know better than that. Recalls are difficult using the SAME GEAR, much less a different compressor of the same model.

And I'm not even going to touch the comment about using a diff comp with similar settings...

I guess once again we'll agree to disagree - I think the points speak for themselves. I'll leave you with this - if a mixer can't mix where they don't need the mastering engineer to finish it, perhaps they're not ready to be mixers...

[masteringhouse]

Originally Posted by bblackwood

Come on, Tom - you know better than that. Recalls are difficult using the SAME GEAR, much less a different compressor of the same model.

And I'm not even going to touch the comment about using a diff comp with similar settings...

Sure, I should have qualified as "reasonably reproduced". But the intent of mastering with stems is to improve upon, not duplicate exactly otherwise there is no point in going through the extra hoops. When doing an up/down mix the stereo bus compression would likely change anyway.

Originally Posted by bblackwood

I'll leave you with this - if a mixer can't mix where they don't need the mastering engineer to finish it, perhaps they're not ready to be mixers...

We agree, though I will add that the "perfect mix" is in the ear of the beholder and can change. Clients do change their minds after living with a mix for a while, and environmental differences between the studio and mastering environment reveal different aspects about a mix that can totally change your perspective. Where the bass or vocals seem just right in one they may be overbearing in another. Where would you like the final decision to be made, before committing to the final phase of production or having one last chance to make corrections during?

Brad we don't really disagree much on this, in the hands of a great mix engineer mixing is a performance that shouldn't be tampered with much like the conductor of a great orchestra. In a home recording forum, well ...

[tarnationsauce2]

99% of the time I think a stereo mix to be mastered as normal will yield superb results. Stem mastering will not pull much else out of the mix.

However I have ran across bands that want loud clicky bass drums (at the same time as really loud guitars), and the best way I found to keep the clicks in is by steming the bass drum track.
On a stereo mix the multiband will engage smushing the bass drum click when the the guitar or cymbals come in since they are in the same frequency range as the bass drum clicks. So what would happen is during the quiet parts of the song, there would be a lot of definition and click to the bass drum, but when it got loud and everything comes in, the bass drum click gets buried.
But with a stemed bass drum, the click stays very pronounced.

The dynamics signal chain will be something along these lines:

Stereo mix (minus bassdrum) into a multiband compressor (no limiter, or only a small amount).
Bass drum into a multiband compressor (no limiter, or only a small amount).
Mix the compressed Stereo mix and bassdrum together and apply some limiting to keep everything in check.

Yes that does indeed require a bit of mixing in the mastering house. But it's probably the most powerful way to get super loud click bassdrums. Not that I prefer the bassdrum to be so loud, but some people do.